用自动调优让LLM在不同GPU上高效运行，无需改代码。
GPU Performance Portability needs Autotuning
- 结合即时编译与参数自动调优，提升性能兼容性。
- 探索15倍更多参数配置，最高提速230%且代码量缩小70倍。
- 适合想跨GPU部署模型的研究者和开发者。
随着大语言模型（LLM）复杂度增加，达到顶尖性能需要算法、软件与硬件的紧密协同设计。当前对单一主流平台的依赖限制了模型的可移植性，导致厂商锁定，并提高新硬件进入门槛。本文提出将即时编译（JIT）与全面的内核参数自动调优相结合，实现无需代码修改的高性能可移植LLM推理。针对关键性能内核，实验表明该方法可探索最多15倍的参数配置组合，生成显著更丰富的代码形态，在多个维度上实现多样性提升；即使在未优化的情况下，其性能仍可超越厂商优化版本达230%，同时内核代码规模减少70倍，彻底消除手动优化需求。结果表明，自动调优是实现跨厂商GPU模型可移植性的可行路径。
原文摘要 · Abstract (English)
As LLMs grow in complexity, achieving state-of-the-art performance requires tight co-design across algorithms, software, and hardware. Today's reliance on a single dominant platform limits portability, creates vendor lock-in, and raises barriers for new AI hardware. In this work, we make the case for combining just-in-time (JIT) compilation with comprehensive kernel parameter autotuning to enable portable LLM inference with state-of-the-art performance without code changes. Focusing on performance-critical LLM kernels, we demonstrate that this approach explores up to 15x more kernel parameter configurations, produces significantly more diverse code across multiple dimensions, and even outperforms vendor-optimized implementations by up to 230%, all while reducing kernel code size by 70x and eliminating manual code optimizations. Our results highlight autotuning as a promising path to unlocking model portability across GPU vendors.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。